Vercel breach after OAuth token leak in Context AI tool

Vercel disclosed that it was breached through a third‑party AI tool, Context[.]ai, after a compromised OAuth token belonging to a Vercel employee who signed up for Context’s AI Office Suite using their Vercel Google Workspace account granted broad permissions.

According to the company’s security bulletin, this enabled the attacker to access some Vercel environments and environment variables that were not marked as “sensitive.” The incident is being investigated with Mandiant and other security firms, and Vercel has contacted customers believed to be at risk and advised credential rotation for a limited subset affected downstream.

No evidence has been presented that sensitive variables were read, and Context stated its own advisory about related activity; the overall assessment emphasises the risk posed by AI data security and unsanctioned tools.

A darkReading debate quoted that the breach is “a threat actor [allegedly ShinyHunters]” selling the incident for $2 million, and while that claim is noted, Dark Reading stresses that OAuth tokens remain a principal attack surface, with attackers exploiting vendor access to move laterally downstream. 20 April 2026.